Find the fault. Do not reboot it into the next hour.
A site 502, SSH refused, mail deferred, load at the ceiling. We read the log and the disk before restarting the service that will fail again.
Disk, then the service, then the application
Ping, SSH, the provider console. If SSH is dead and the disk is full, the path is rescue, not a ticket that says “please reboot”.
NGINX, PHP-FPM, MariaDB, Exim or Postfix. The error is in that unit’s log. A 502 with a healthy web server is usually PHP or the upstream.
An update, a deploy, a full disk, a certificate. If nothing changed, the next suspect is capacity or a job that filled the volume.
The level matches how down it is
A broken config in business hours is €55/hour, minimum 1 hour. A production site or mail system impaired is priority at €75. Down, hacked, or disk full is emergency at €95, subject to availability. Out of hours is €125, and only if someone is free.
The cause stays in the note
You get what failed and what was changed. A monthly plan is optional. If the same fault will return because there is no backup or no monitoring, we say that. We do not open a retainer to finish the ticket.
Something failed, and a reboot did not hold?
Tell us the provider and the error. No root password in the first message.